TWI784051B - Semiconductor device attaching method - Google Patents

Abstract

The present invention relates to an attaching method of a semiconductor material attaching device that can quickly and accurately correct a positional error of an attaching position during attaching a semiconductor material. The present invention is characterized in that the vision or the table is moved at intervals calculated based on the matrix information of the attachment area to which the material is to be attached detected within the viewing angle of the vision unit, and it is acquired that each target attachment area is at a position different from each other multiple images and determine the location of the target attachment area from these images.

Description

Semiconductor Material Attachment Methods

The present invention relates to semiconductor material attachment methods. More specifically, the present invention relates to a method for attaching a semiconductor device (Semiconductor Device Attaching Method) that improves inspection accuracy by accurately detecting the attachment position of the semiconductor material.

In the semiconductor material attachment device, the individualized semiconductor material must first accurately grasp the preset attachment position of the attachment object in order to attach the attachment object for the subsequent process.

Such attachment means may be adhesive means for bonding a plurality of semiconductor materials to a substrate, or attachment means for attaching to a tape for other subsequent processes. Also, it may be an attachment means for attaching semiconductor material to the tape to accommodate spherical surfaces of semiconductor material in the tape perforated by the semiconductor material for electromagnetic compatibility (EMI) sputtering for shielding electromagnetic waves.

Especially, in the attaching device for electromagnetic compatibility sputtering, when attaching the semiconductor material on the tape, it is necessary to accurately detect the position of the hole of the tape and attach it at the correct position so that accommodate semiconductors The spherical surface of the material (bumps), thus protecting the bumps from the electromagnetic shielding material. If the semiconductor material is not attached to the correct position of the hole, the leaked portion will also be sputtered to the bump of the semiconductor material, thus adversely affecting the electrical properties of the semiconductor material.

Therefore, due to the optical offset value (X axis, Y axis, Z axis) caused by the vision camera used to detect the attachment table or position, the position error on the plane or the offset on the θ axis, etc. The error is reflected in the accuracy, therefore, it is very important to accurately detect multiple attachment positions on the attachment object for the attachment process.

In order to solve these problems, when each attachment location is photographed several times through the vision unit above the attachment location for the accuracy judgment of an attachment location, much time can only be spent on accuracy inspection.

On the other hand, in recent years, the performance of the semiconductor process has been improved, and the number of high-speed, high-resolution cameras has increased, and the size of semiconductor materials has gradually become smaller. Therefore, the number of materials entering the field of view (FOV, field of view) has increased.

Therefore, although all materials entering into the field of view are inspected one by one in order to increase productivity, the mechanical error value of the attached table and vision camera has to be reflected in the accuracy.

In order to solve the above problems, the present invention provides a method for attaching a semiconductor material attachment device that can quickly and accurately detect the attachment position of the semiconductor material.

In order to solve the above-mentioned problems, the present invention can provide a semiconductor material attachment method, the above-mentioned semiconductor material attachment method is a circuit substrate having a bonding area formed with a plurality of semiconductor materials for bonding, a workbench for placing the above-mentioned circuit substrate, and A method for attaching a semiconductor material attaching device of a visual unit for photographing the bonding area of the above-mentioned circuit substrate, the above-mentioned semiconductor material attaching method includes: using the above-mentioned visual unit, photographing the bonding area of the above-mentioned semiconductor material with a single lens (shot). The first photographing step of the combined target bonding area and a plurality of adjacent bonding areas; The step of transferring the visual unit or the workbench at the interval calculated by the matrix information of the bonding area of the angle of view; in the state of transferring the vision unit or the workbench at the above-mentioned calculated interval, using the above-mentioned vision unit to photograph the above-mentioned target bonding The second photographing step of the area; by repeating the above-mentioned transfer step and the second photographing step multiple times, a step of obtaining multiple images of the above-mentioned target bonding area in different positions within the viewing angle of the above-mentioned visual unit; and from the above-mentioned A step of determining the position of the target bonding area by acquiring a plurality of images of the target bonding area.

And, in order to solve the above-mentioned problems, the present invention can provide a semiconductor material attaching method, the above-mentioned semiconductor material attaching method has a through hole formed with a plurality of bumps for accommodating the semiconductor material and is used for the sputtering process of the above-mentioned semiconductor material A method of attaching a semiconductor material attachment device to a tape of a template, a table for placing said tape, and a vision unit for photographing a through-hole of said tape, said method of attaching semiconductor material comprising: using said vision unit, Take a single-shot shot of the above-mentioned semiconductor material to be contained The first shooting step of the target through hole of the bump and a plurality of adjacent through holes; after shooting the above target through hole, in order to make the target through hole enter other positions in the visual angle of the above visual unit according to the viewing angle of entering the visual unit The step of transferring the vision unit or the workbench at the interval calculated by the matrix information of the through hole; in the state of transferring the vision unit or the workbench at the above-mentioned calculated interval, using the above vision unit to photograph the second of the above-mentioned target through-hole A photographing step; by repeating the above-mentioned transfer step and the second photographing step multiple times, a step of obtaining multiple images of the above-mentioned target through holes in mutually different positions within the viewing angle of the above-mentioned visual unit; and from the obtained multiple above-mentioned A step of judging the position of the above-mentioned target through-hole by using the image of the target through-hole.

In this case, the above-mentioned first photographing step and the second photographing step repeat photographing at the respective positions multiple times, and the position can be determined by using an average value of multiple position values acquired through repeated photographing.

And, when a specific abnormal position value is found from the position value of the first photographing step and a plurality of position values obtained by repeating a plurality of second photographing steps, the corresponding data is filtered, and the position value of the first photographing step and When a plurality of data among the plurality of position values obtained by repeating the second shooting step for several times have different deviations, recalibration may be performed or the corresponding position value may be regarded as bad.

And, the above-mentioned bonding area entering the viewing angle of the above-mentioned visual unit can be formed into M rows×N columns, and the above-mentioned M and N can be integers. In the above-mentioned second shooting step, when the above-mentioned M is an even number, the M/ 2-column intervals, while shooting the above-mentioned target bonding area, when the above-mentioned M is an odd number, the above-mentioned target bonding area can be photographed while moving (M+1)/2 columns, when the above-mentioned N When it is an even number, the above-mentioned target bonding area can be photographed while moving N/2 line intervals, and when the above-mentioned N is an odd number, it can be photographed while moving (N+1)/2 line intervals.

Moreover, the above-mentioned through hole entering the viewing angle of the above-mentioned visual unit can be formed as M rows×N columns, and the above-mentioned M and N can be integers. In the above-mentioned second shooting step, when the above-mentioned M is an even number, it can be moved by M/2 Column spacing, while photographing the above-mentioned target through-hole, when the above-mentioned M is an odd number, you can move (M+1)/2 columns while photographing the above-mentioned target through-hole, and when the above-mentioned N is an odd number, you can move N/2 rows interval, while photographing the above-mentioned target through-holes, when the above-mentioned N is an odd number, the above-mentioned target through-holes can be photographed while moving (N+1)/2 row intervals.

In addition, the target bonding area may be photographed by the vision unit while moving with the vision unit or the stage at a pitch interval.

In addition, the vision unit may be used to photograph the target through hole while the vision unit or the table is moving by a pitch.

In addition, in the step of acquiring a plurality of images in which the above-mentioned target bonding area is at different positions, the above-mentioned target bonding area is located at the upper left, upper right, and lower left with the center of the visual unit as a reference so that the target adhesive area can be obtained. and the image at the lower right, the distance can be captured by the vision unit to obtain the image.

And, in the step of acquiring a plurality of images in which the above-mentioned target through-holes are at different positions from each other, in order to obtain the above-mentioned target through-hole Using the center of the vision unit as a reference, the intervals are calculated based on the images located at the upper left, upper right, lower left and lower right, and the vision unit or the workbench is moved while the vision unit is used to capture images.

In addition, the above-mentioned template has a plurality of through holes larger than the through holes of the above-mentioned tape in positions corresponding to the through holes formed in the above-mentioned tape, and in the above-mentioned visual unit photographing step, it is possible to extract the surrounding area of the through-hole of the above-mentioned template and The image of the periphery of the through-hole of the above-mentioned tape is used to obtain the tolerance between the through-hole of the above-mentioned template and the through-hole of the above-mentioned tape, and it is confirmed whether the obtained tolerance is within the initial setting range.

In addition, the above template has a plurality of through holes larger than the through holes of the above-mentioned tape at positions corresponding to the through holes formed in the above-mentioned tape. The tolerance between the via hole and the via hole of the above-mentioned strip is compared with the position of the bump of the above-mentioned semiconductor material to check the adhesion state of the semiconductor material.

According to the attaching method of the semiconductor material attaching device of the present invention, even if the semiconductor material such as a semiconductor wafer and the attachment position (circuit board, tape) to which the semiconductor material is attached are miniaturized, the attached The position error of the contact position can be improved, and the accuracy can be improved.

Furthermore, according to the semiconductor material attachment method of the present invention, a plurality of attachment positions are arranged within the viewing angle of the visual unit, and the positions of the target attachment positions are arranged at different positions within the viewing angle, through which the photographed Multiple images of the target are used to judge the position error of the target attachment position, so the accuracy of the judgment of the position error can be improved.

And, according to the semiconductor material attachment method of the present invention, the optimal pitch interval calculated according to the matrix information of the attached object detected in the viewing angle is moved, and through the overlapped images acquired in each inspection , to obtain the image information of a target attachment position detected in the upper left, lower right, lower left and lower right positions based on the center of vision. Therefore, not only the multi-lens effect for a target attachment position can be obtained, It can also eliminate mechanical, imaging, and positional defects during visual inspection, thereby reducing defect images and obtaining reliable image information.

And, according to the semiconductor material attaching method of the present invention, in the case where the attached objects (materials) detected within the viewing angle increase due to the small size of the semiconductor material, the matrix of the attached objects detected within the viewing angle is also Information to calculate the optimal pitch interval to move the vision, so that the visual inspection position is different, so that the visual inspection speed can be shortened to increase the unit per hour (UPH, nit per hour).

And, check all the attachment positions that can be detected within the viewing angle, and detect through the detected target attachment positions between overlapping lenses, so that a multi-lens effect for one material can be realized, and detection through multiple lenses can be achieved. By calculating the average value of the obtained multiple positions and calculating the accurate position value, the mechanical error value can be confirmed and the influence of the error value can be eliminated.

Moreover, when the shooting value at a specific position is not good among the captured images of the detected multiple lenses, the corresponding data can be filtered and used, or when the value of the specific position is repeated abnormally, the image plane at the specific position can be judged (illuminance) is problematic, so accurate position values are calculated by reflecting that problem, or when all images from multiple lenses have different deviations , recalibrate the corresponding position or consider it as bad and exclude it in the subsequent process, so that bad phenomena can be prevented in advance.

Moreover, according to other embodiments of the attachment method of the semiconductor material attachment device of the present invention, material deviation inspection and post bonding inspection (PBI, Post Bonding Inspection) can also be performed.

100: Attachment (belt)

100':wafer

110: Attachment position (through hole)

110': bonding area

tap: target attachment position (target glue position)

200, 200': visual unit

fov: angle of view

sp: semiconductor wafer

FIG. 1 is a plan view showing an attachment article having a plurality of attachment positions for attaching semiconductor materials in a first embodiment of the present invention.

Fig. 2 is a diagram showing a state in which the vision unit photographs the attachment position before the attachment process in the first embodiment of the present invention.

Fig. 3 shows the relative movement of the attachment workbench for placing the vision unit and the attached objects in the first embodiment of the present invention, and the target attachment positions are arranged in different positions and photographed including the target The process of attaching multiple images of a location.

FIG. 4 is an image showing overlapping target attachment positions in a plurality of images captured by the vision unit of FIG. 3 in the first embodiment of the present invention.

5 is a cross-sectional view showing an attachment hole of a sputtering member to which a sputtering tape used in a sputtering process is attached and a solder ball of a semiconductor material to be sputtered is attached to the attachment hole in the first embodiment of the present invention. It is a cross-sectional view of a state of a semiconductor wafer of a package-in-array (BGA) system.

FIG. 6 is a diagram showing a state in which a vision unit of a semiconductor material attaching device according to a second embodiment of the present invention photographs an adhesive region before an attaching process.

7 is a diagram showing a state in which the wafer stage for placing the vision unit and the wafer is relatively transferred in the semiconductor material attaching device according to the second embodiment of the present invention, and the target bonding area is arranged at different positions from each other. The process of taking multiple images including the target bonded area.

FIG. 8 is an image showing overlapping target bonding regions in a plurality of images captured by the vision unit of FIG. 7 in a second embodiment of the present invention.

9 is an image in which the vision unit of the semiconductor material attaching device according to the second embodiment of the present invention photographs and overlaps the bonding area while moving in the X-axis direction at calculated pitch intervals before the attaching process .

Hereinafter, several preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, the invention is not limited to the embodiments described here, but can also be embodied in different ways. Rather, various embodiments described herein are provided so that the disclosure will be thorough and complete, and to fully convey the concept of the invention to those of ordinary skill in the art to which the invention pertains. Throughout the specification, the same reference numerals denote the same structural elements.

FIG. 1 is a view showing a semiconductor material attachment device (attaching device) with an attachment unit according to a first embodiment of the present invention through an attachment object 100 having a plurality of attachment positions 110 through which semiconductor materials are picked up and attached. A plan view, FIG. 2 is a state showing a photograph taken by the vision unit 200 in the first embodiment of the present invention for judging the position error of the attachment position 110 before the attachment process.

A semiconductor material attaching device used in the semiconductor material attaching method of the present invention, for example, a bonding device for bonding a semiconductor material of an adhesive object in a substrate, or an attaching device for attaching a semiconductor material on a tape for a subsequent process . In addition, it may also be a semiconductor material attaching device used for bonding or attaching a semiconductor material as a sputtering object on the sputtering member S to which the sputtering tape t is attached for the sputtering process, but is not limited thereto. It is applicable to the case where the semiconductor material is attached to the attachment object 100 in the semiconductor process.

The semiconductor material attaching method of the present invention can be roughly divided into a first attaching method for adhering the semiconductor material to a substrate or a wafer and a second attaching method for attaching the semiconductor material to a tape.

First, the semiconductor material attaching device used in the first attaching method of the present invention includes: a circuit substrate formed with a bonding area for bonding a plurality of semiconductor materials; a table for placing the above-mentioned circuit substrate; and The vision unit, which shoots the target bonding area to be bonded of the above-mentioned semiconductor material and adjacent multiple bonding areas with a single shot (shot), and shoots multiple times, so that the above-mentioned target bonding area passes through multiple shots in other The above-mentioned semiconductor material attachment device is characterized in that the above-mentioned circuit substrate and the above-mentioned visual unit can be moved by relative movement, and the above-mentioned visual unit moves and photographs the bonded area at intervals of a pitch. The distance interval is calculated based on the matrix information of the attached object detected within the view fov.

Here, the circuit substrate may be a rectangular substrate or may be a wafer.

In the present invention, the circuit substrate and the vision unit can be moved and photographed by relative motion. In this case, in terms of the relative movement between the circuit substrate and the above-mentioned visual unit, in the state where the circuit substrate can be moved in the X-axis and Y-axis directions, the visual unit can be a fixed type, otherwise, the circuit substrate can be a Fixed type, and the visual unit can be set in the direction of X-axis and Y-axis, and the circuit board can also be moved in the direction of X-axis (or Y-axis), and the visual unit can be moved in the direction of Y-axis (or X-axis) It is set in such a way that the circuit substrate and the vision unit are photographed as they move relative to their respective uniaxial directions.

This structure can be properly deformed and utilized according to the structure of the staff and equipment.

In the present invention, detecting the target adhesive area at other positions via multiple lenses by shooting multiple times may refer to the target adhesive area (attachment area) that enters the visual angle of the vision unit, and the target adhesive area is visually recognized. The center is located at the upper left, lower left, lower right and lower right respectively, that is, through the relative movement of the visual unit and the circuit substrate, multiple shots are taken, so that relative to the same target bonding area, the target bonding area is located at the center of the vision Near the upper left (upper left), lower right (upper right), lower left (lower left), and lower right (lower right) near the center, the position of the target bonding area can be accurately detected through these images.

The semiconductor material attachment device used in the second attachment method of the present invention comprises: a tape formed with a plurality of through holes for receiving bumps of the semiconductor material, and attached to a template for a sputtering process of the semiconductor material ; a table for placing the above-mentioned belt; and a vision unit for shooting the above with a single lens The target through-hole to be inspected and the adjacent multiple through-holes among the multiple through-holes are photographed multiple times, so that the target through-hole is detected at other positions through multiple lenses, and the semiconductor material attachment device It is characterized in that the above-mentioned workbench and the above-mentioned vision unit are arranged in a relatively movable manner, and the above-mentioned vision unit photographs the above-mentioned through hole at a pitch interval, and the pitch interval is based on the matrix information of the attached object detected within the viewing angle to calculate.

In the semiconductor material attaching device used in the first attaching method and the second attaching method of the present invention, when the vision unit is photographed multiple times through a plurality of lenses to detect the target bonding area or the target through hole at other positions , but also can repeat shooting more than two times at each position. When repeated shooting at one position more than twice, the influence of vibration applied to the device due to device driving or external factors can be eliminated, so that more reliable and accurate position values can be obtained.

In recent years, the size of semiconductor materials such as semiconductor wafers has been miniaturized, and in the case of attaching this to the attachment object 100 (circuit board or tape), it is also required to attach due to the miniaturization of the size of the terminals of the substrate, etc. The precision of the process.

As mentioned above, for the accuracy of the attaching process, the semiconductor material attaching device of the semiconductor material usually has a function for extracting the target attachment position 110 (target bonding area or target through hole) to be attached by photographing the semiconductor material, including the The vision unit 200 of the image of the attachment position 110, before attaching the semiconductor material to the attachment object 100, the control part of the semiconductor material attachment device moves the vision unit 200 or the attachment workbench to a preset standard value and photographs the target attachment location. At this time, since the vision unit or the X Axis, Y-axis and Z-axis offset (straightness, flatness, rolling, pitching, yawing), even if the preset standard value is moved, it is possible Errors occur within the viewing angle of the visual unit according to each working position, and may also cause visual detection errors and incorrect bonding of semiconductor materials due to mechanical errors or external environment variables.

Therefore, when inspecting semiconductor materials that enter the viewing angle, different error values may occur depending on the position of the mechanical error value inspection or the detection area within the viewing angle area. Therefore, in the present invention, the average error can be ensured through multi-position inspection. value.

Therefore, if the average or compensation value of a plurality of offsets detected for each position is obtained, the influence of errors occurring in mechanical parts can be reduced and accuracy can be ensured.

The inspection method of the present invention is particularly useful when relative correction cannot be performed because there is no reference or fiducial during visual inspection. By repeatedly obtaining the position error of the target attachment position, calculating the compensation value of the error value detected for each position, and using the obtained compensation value, the error generated in the mechanical part can be eliminated.

Here, the so-called target attachment position refers to the target attachment position where the semiconductor material passes through the attachment unit in a specific order among the multiple attachment positions where the semiconductor material needs to be attached. The multiple attachment positions can potentially be target attachment positions. Location. In the present invention, the target attachment location may be referred to as a target bonding area or a target via.

Recently, due to the improvement of the graphic element, resolution, or viewing angle of the lens of the image device constituting the vision unit 200, etc., the trend of gradually reducing the size of the semiconductor material, therefore, the number of materials entering the viewing angle is increasing, which is continuing. Developing methods to accurately detect wide areas using vision.

Conventionally, when the vision unit 200 is used to photograph the target attachment position in order to determine the positional error of the attachment position for attaching the semiconductor material, due to variations in the quality of the photographed results, etc. It is a method of judging the position error of the target attachment position by shooting multiple times with the attachment position arranged in the center. However, factors such as the direction of light or shadows are not easy to change during the shooting process. Therefore, even if multiple shots are taken Even in the same target attachment position, an image for accurately determining a positional error cannot be obtained. Also, as the respective target attachment positions are photographed multiple times, considerable time is consumed in checking many target attachment positions multiple times.

However, the semiconductor material attachment device of the present invention inspects all semiconductor materials that enter the viewing angle fov of the vision unit, and can further improve the accuracy and reliability by utilizing the overlapping inspection multi-lenses in the state of moving each pitch sex. That is, the present invention photographs the target attachment position to be inspected and adjacent plural attachment positions with a single lens, and photographs a plurality of times as each pitch is moved, so that it is possible to attach the target at other positions via a plurality of lenses. The attachment position is located on the left upper part, right upper part, left lower part and right lower part centered on the visual center to detect the target attachment position, so that when the shooting surface and the working position are changed, the same detection error can be eliminated, so that it can be realized Check for reliability.

Therefore, in order to accurately determine the position error of the target attachment position 110, the semiconductor material attachment device of the present invention may include a vision unit 200, which shoots the target attachment to be inspected in the above-mentioned attachment object 100 with a single lens. The attachment position and adjacent multiple target attachment positions are photographed multiple times, so that the target attachment position is detected at other positions via multiple lenses. The above-mentioned visual unit movement is equivalent to the The distance of the pitch calculated by the matrix information is photographed so that the above-mentioned target attachment position is detected at other positions through multiple lenses, so that an overlapping image of the attachment position detected between the lenses can be acquired, thereby enabling Accurate position values are calculated between target attachment positions.

Hereinafter, as a semiconductor material attaching device according to a first embodiment of the present invention, it is explained with reference to FIGS. 1 to 5 . In the first embodiment, the second attachment means for attaching the semiconductor material to the tape will be described as an example.

The aforementioned attached object can be a circuit substrate or a wafer, and can also be a ring frame with a tape attached, or a sputtering part or a sputtering part with a sputtering tape for a sputtering process of a semiconductor material attached, but in FIG. 1 may be named attachment object 100. The attachment locations 110 can be arranged in a plurality of columns and rows in the tape. Here, the tape (attachment object) is a sputtering tape used in a sputtering process of semiconductor materials, and is attached to a sputtering member or template for supporting the tape, and can be supplied in a state placed on a tape table .

As shown in the figure, four attachments are photographed in the viewing angle fov of the vision unit 200 of the semiconductor material attachment device according to the first embodiment of the present invention. Attachment positions 110, four attachment positions 110 are captured together in one shot image.

In the above attachment position, the frame portion of the semiconductor material is attached to the tape in a state where the spherical surface (bump) of the semiconductor material is accommodated. The attachment positions 110 photographed through the vision unit can be arranged in four within the viewing angle of the vision unit 200 as shown in FIG. method to move each pitch and take a picture of the belt attachment location, inspecting it by overlapping the attachment locations between the shots, which means that one attachment location can be detected in other locations in a way that goes through multiple shots shoot.

That is, when the image of the target attachment position is inaccurate due to factors such as the presence of foreign objects, interference, the direction of light or shadows, and vibrations generated in the device, within the viewing angle shown in Fig. 2 After the method of changing the position of the target attachment position is transferred, it is photographed to accurately determine the position error information of the target attachment position.

For example, determining the position error by averaging the position error of the target attachment position through multiple images, or treating images with particularly large error values as errors generated during the shooting process and excluding or filtering them, Only use the excellent data as the judgment data of the position error.

In the first embodiment of the present invention, the visual angle of the vision unit can detect images of 4 attachment positions with a single lens of 2 rows×2 columns. The matrix information of the attachment position of the viewing angle into the vision unit selectively, the shooting interval can be different.

Fig. 3 shows the vision unit 200 of the semiconductor material attachment device according to the first embodiment of the present invention and the table transfer for placing the attachment object 100, and the target attachment position tap formed on the tape is arranged in different places. In the state of the position, the process of taking a plurality of images including the target attachment position tap.

As a method of changing the position of the target attachment position tap within the viewing angle fov of the vision unit 200 , at least one of the table and the vision unit 200 may be transportable on the X-Y plane.

One mechanism in the above-mentioned workbench and the above-mentioned visual unit 200 can be configured to be transferred on the X-Y plane, but it is also possible that the above-mentioned visual unit 200 is transferred to the X-axis (or Y-axis) direction, and the above-mentioned workbench is transferred to the Y-axis (or Y-axis) direction. X-axis) direction, so that the visual unit and the table can move relative to each other.

In the embodiment shown in FIG. 3 , the case where the vision unit 200 can be moved in the X-axis direction and the attachment table on which the attached object 100 is placed can be moved in the Y-axis direction is illustrated as an example.

FIG. 3 shows that in the case where the above-mentioned target attachment position tap is photographed as 2 rows and 2 columns in the viewing angle fov of the vision unit 200, as the vision unit 200 or the workbench is transferred, the target attachment position tap The process of changing the position and taking pictures, FIG. 4 is an image showing the overlapping of multiple images taken in FIG. 3 with the above-mentioned attachment position 110 as the center.

(a) of FIG. 3 shows the state where the target attachment position tap is arranged at the lower right side under the initial condition, and (b) of FIG. The state where the transfer is performed on the right side in the X-axis direction and the above-mentioned target attachment position tap is shifted to the lower left side, Fig. 3 (c) In the state shown in (b) of Fig. 3, the attaching workbench is moved downward in the Y-axis direction and the above-mentioned target attachment position tap is shifted to the upper left side, and (d) of Fig. 3 shows In the state shown in (c) of FIG. 3 , the visual unit 200 is moved to the left in the X-axis direction and the above-mentioned target attachment position tap is shifted to the upper right.

That is, when this is represented by coordinates, the effect that can be obtained is that when the position coordinates of the left lower part in the viewing angle are assumed to be (1, 1), the above-mentioned target attachment position tap is shown in Fig. 3(a), Fig. The coordinates of 3(b), 3(c) and 3(d) move to the four coordinates of (1, 2), (1, 1), (2, 1) and (2, 2).

Referring to Fig. 2 to Fig. 4 as the first embodiment of the present invention, when the viewing angle of the visual unit is 2 rows × 2 columns, shooting is performed every time one pitch is moved, and each attachment position overlaps to have each Multi-shot effect of 4 shots at the target attachment position. If the viewing angle of the vision unit is 3 rows x 3 columns due to the smaller size of the material to be attached, each target can be attached by overlapping each attachment position when shooting by moving 1 pitch Multi-camera effect with position 9 times.

As mentioned above, in the present invention, one target attachment position tap is arranged at different positions within one viewing angle, preferably at four positions (upper left, lower right, lower left) based on the visual center. and lower right), so when judging the position error of the target attachment position by acquiring multiple images as a reference, the accuracy of judging the position error can be improved.

That is, when multiple images are captured with the attachment position 110 arranged at the same position within the angle of view, it is also likely to be affected by foreign objects, light quantity, direction of light, shadows, etc., but when the visual unit 200 When shooting together with the attachment position 110 within the angle of view, as shown in FIG. 3 , the position of the target attachment position tap can be changed within the angle of view of the vision unit 200, and the target can be judged through various images including the target attachment position tap A more accurate location of the attachment location.

As shown in FIG. 4 , in the four captured images, the relative positions of the above-mentioned target attachment positions tap are different from each other, but the positions of the attachment positions 110 arranged at different positions are photographed to compare and analyze each image. The error reflected in the image, and the position error can be corrected and attached during the process of attaching the semiconductor material.

Therefore, in the above-mentioned example, the vision unit 200 can capture the attachment positions 110 in 2 rows×2 columns with one image, and the vision unit can acquire the state that the attachment positions 110 are arranged at each of the 4 positions in the 2 rows×2 columns. 4 images, and compare their positions with the set standard value, so that 4 position error values can be obtained. Part of the shooting values of their position error values, especially when the value obtained at a specific position is out of the error range or an error occurs, it is considered to be due to external factors such as the drive of the device or the unevenness of the detection surface or particles when shooting at the corresponding position. For the generated error, the corresponding position value can be used after filtering. If deviations occur in all four position error values, re-check or consider that there is a problem in the corresponding position, and omit the subsequent semiconductor material attachment in the corresponding position, so that material waste and defects can be minimized.

To sum up, the method for attaching a semiconductor material according to the first embodiment of the present invention includes: an attachment position (adhesive area) where the semiconductor material is to be attached and an attachment position (adhesive area) adjacent to the above-mentioned attachment position (adhesive area) A visual unit photographing step in which the position (adhesive area) is arranged within the viewing angle of the visual unit, and a visual unit imaging step; a target of transferring the attached position (adhesive area) to another position within the visual unit’s viewing angle a transfer step; a position error judging step of judging the position error of the above-mentioned attachment position (adhesive area) through the image captured in the above-mentioned visual unit photographing step; according to the judging result of the above-mentioned position error judging step, the semiconductor material is carried out The step of attaching the semiconductor material, the step of photographing by the vision unit and the step of transferring the object are repeated multiple times, and the step of determining the position error uses a plurality of images obtained based on the plurality of images captured in the step of photographing by the vision unit. The method of judging the position of the attachment position (adhesive area) by averaging the position values of the image, thereby improving the accuracy of position judgment and attaching the semiconductor material accurately.

At this time, the average of multiple position values obtained by repeating the visual unit shooting steps is calculated, and can be set as the position value, and the position error value can be calculated through each position value, and the position compensation can also be obtained from this value.

On the other hand, when a specific abnormal position value is found among the multiple position values obtained by repeating the visual unit shooting steps, the corresponding value can be filtered, and in the case of repeated abnormalities in the specific position value, it can be known that the specific position There is a problem with the image surface (illuminance), so by reflecting this to calculate an accurate position value, if there is a deviation in multiple values, you can recheck or treat the corresponding position value as bad.

5 is a cross-sectional view showing an attachment hole of a sputtering member S to which a sputtering tape t for a sputtering process is attached according to a first embodiment of the present invention and through the attachment of a semiconductor material according to the first embodiment of the present invention. A cross-sectional view of a semiconductor wafer of a ball array packaging system in which a semiconductor material to be sputtered is attached to the attachment hole of the device.

As mentioned above, the above-mentioned semiconductor material is a semiconductor material of a ball grid array package (BGA, Ball Grid Array) type with ball electrodes on the bottom surface, and the above-mentioned attached object 100 is a sputtering tape t attached to a sputtering process for the semiconductor material. The sputtering component S, the attachment position 110 may be a plurality of holes formed to penetrate the ball electrode surface of the semiconductor material formed on the sputtering zone t and the sputtering component S together.

The aforementioned sputtering member and the aforementioned sputtering belt t form holes th, holes sh in corresponding positions shown in part (a) of FIG. 5 , and form through holes as target attachment positions 110 . The holes formed in the above-mentioned sputtering part and the above-mentioned sputtering belt t are formed in corresponding positions, and form through holes, but are prevented from being formed in the sputtering work by making the size of the holes formed in the sputtering belt t smaller. The slits of the holes leak and sputter on the spherical surface of the material and prevent possible contamination of the sputtering components by the leaked sputter deposition agent.

During the sputtering process of semiconductor materials in the BGA (Ball Grid Array) method, in order to prevent the ball electrodes or ball electrode surfaces on the lower surface of the semiconductor wafer from being sputtered, the sputtering tape coated with an adhesive substance A through hole is formed in the center, and the bottom frame of the conductor wafer is attached to the periphery of the through hole, and the ball electrode can be exposed downward through the through hole.

That is, as shown in part (b) of Figure 5, preferably, the size of the above-mentioned semiconductor wafer is larger than the size of the through hole of the sputtering zone, so that the spherical bumps on the lower surface of the semiconductor wafer are not sputtered, and the frame of the semiconductor wafer Parts tend to attach in the sputter band. A plurality of through-holes slightly larger than the belt holes are provided in the sputtering part at positions corresponding to the through-holes of the belt, and both the through-holes of the sputtering part and the through-holes of the belt are rectangular openings. In addition, the thickness of the sputtering member is substantially the same as or slightly thicker than the thickness of the bumps of the semiconductor wafer.

On the other hand, in the present invention, when the semiconductor wafer is attached in the hole of the tape, the hole of the tape cannot be detected because the size of the semiconductor wafer is larger than that of the tape. For this reason, after the semiconductor material is attached, when the attached material is overturned, both the frame of the template and the frame of the sputtering tape can be detected, so by extracting the periphery of the template and the tape separately, their tolerances can be detected and can be carried out. Retest. In addition, the bumps of the semiconductor wafer are inspected from the periphery of the template, so that automatic post-bonding inspection (PBI, Post Bonding Inspection) after bonding can also be realized.

In the step of photographing by the vision unit, the tolerance between the through hole of the above template and the through hole of the tape is obtained by extracting images of the periphery of the through hole of the above template and the periphery of the above mentioned through hole, and the obtained tolerance is confirmed. Is it within the initial setting range. When it is within the range of the initial setting range, it can be judged whether the band via hole is suitable among the via holes of the template. If it deviates from the initial setting range, the band via hole among the via holes of the sample is defective or the visual image is wrong, so check again. After the re-inspection, if the tolerance of the through hole and the through hole of the sample is not suitable, it is regarded as defective, and the semiconductor material is not attached to the above through hole.

And, after the above-mentioned semiconductor material is attached to the through-hole of the above-mentioned tape, the attachment state of the semiconductor material can be carried out by comparing the tolerance between the through-hole of the above-mentioned model and the through-hole of the above-mentioned tape with the position of the ball electrode of the above-mentioned semiconductor material. Automatic inspection after welding (PBI, post bonding inspection). When performing automatic post-weld inspection or pre-extracting the periphery of the template and tape, the brightness of the inspection can be adjusted by using composite lighting, so that a more vivid image can be obtained.

In the above, the semiconductor material attaching device according to the first embodiment of the present invention is explained with reference to FIGS. Attachment device.

The second embodiment of the present invention is that the size of the material to be bonded is small, so when multiple attachment areas come into view of the vision unit, as described in the first embodiment, the image of each bonding area moves with each section. It takes a lot of inspection time, so it is the best inspection method to move and obtain images according to the pitch interval calculated from the matrix information of the bonded area that is bonded according to the material.

For reference, the attaching method of the second attaching device for attaching the semiconductor material in the tape was described in the first embodiment, but in the second embodiment, the method for attaching the semiconductor material to the wafer is preliminarily exemplified. Attachment method of the first attachment means in 100'.

Also, in the first embodiment, within the viewing angle fov of the visual unit, 2 rows×2 columns of attached objects are detected with a single lens, but in the second embodiment, it is shown that 3 rows×2 columns are detected with a single lens. 3-column case of attached objects. Duplicate content with the first embodiment is omitted.

As shown in FIG. 6 , within the viewing angle fov of the visual unit 200 ′ of the semiconductor material attaching device according to the second embodiment of the present invention, 9 adhesive regions 110 ′ are photographed in 3 rows×3 columns, and one photographed image includes The 9 bonded areas 110' are photographed together.

In the above bonding area 110', the wafer (or substrate) and the wafer (material) are accurately electrically connected to the position where each semiconductor wafer or semiconductor material is attached. Therefore, for fine and accurate bonding, the bonding area ( Accurate location information of attachment location) is particularly important. However, as mentioned above, when each bonding area is photographed several times one by one for the accuracy judgment of the bonding area, the accuracy check can only consume a lot of time, and the accuracy can be determined based on the working position, surface or mechanical error value. There is a defect in the image detected at the position of the target attachment position tap, so all the adhesive regions 110' detectable within the viewing angle are inspected, and the inspected target adhesive regions are overlapped between shots for detection, so that the Multi-shot effect of the targeted bonding area.

7 is a photograph showing a state in which bonding regions formed on a wafer are arranged at positions different from each other within the viewing angle of the vision unit in a semiconductor material attaching device according to a second embodiment of the present invention, including target bonding. The process of combining multiple images of an area.

At this time, at least one of the wafer stage on which the wafer is placed and the vision unit is transferred on the X-Y plane by changing the position of the target bonding area within the viewing angle of the vision unit.

A structure among the wafer workbench and the vision unit 200' can be constituted in a transferable manner on the X-Y plane, and can also be formed with the wafer workbench It is configured in such a way that the visual unit 200' moves in the direction of the X axis (or the Y axis). Thus, the vision unit and wafer stage are movable relative to each other.

In the second embodiment shown in FIG. 7 , the above-mentioned vision unit 200' can be transported to the X-axis and the wafer stage on which the wafer is placed can be transported to the Y-axis.

FIG. 7 shows that when a total of 9 target bonding regions are photographed in 3 rows×3 columns within the viewing angle fov of the vision unit 200′, as the vision unit or the wafer workbench is transferred, the position of the target bonding region changes and In the process of shooting, FIG. 8 is an image showing overlapping the plurality of images shot in FIG. 7 centering on the target bonding area.

Fig. 7(a) shows the state where the target adhesive region is arranged at the lower right under the initial condition.

FIG. 7( b ) shows a state in which the visual unit 200' moves 2 pitches to the right in the X-axis direction and the target bonding area is arranged at the lower left in the state shown in FIG. 7( a ).

FIG. 7( c ) shows a state in which the wafer stage is moved downward by 2 pitches in the Y-axis direction in the state shown in FIG. 7( b ), and the target bonding region is arranged at the upper left.

FIG. 7( d ) shows a state in which the vision unit is moved 2 pitches to the left in the X-axis direction in the state shown in FIG. 7( c ), and the target bonding area is arranged at the upper right.

That is, when this is expressed as coordinates, the following effect can be obtained: the coordinates of the lower left position within the viewing angle are assumed to be located in the target bonding area. In the case of (1, 1), Fig. 7(a) moves to four coordinates of (1, 3), (1, 1), (3, 1) and (3, 3).

In the above, in the first embodiment according to FIG. 1 to FIG. 4 , the visual unit 200 is shown to have a viewing angle of 2 rows×2 columns, so that each via hole overlaps and each bonding area 110' has 4 times The multi-lens effect, according to the second embodiment of FIGS. 6 to 9 , is the case where the viewing angle of the visual unit 200 ′ is 3 rows×3 columns.

In the case of 3 rows × 3 columns, as the viewing angle of the vision unit 200' moves every 2 pitches and shoots, similarly, each bonding area 110' overlaps and each target bonding area has the same 4 times of multi-shot Effect. In this case, the so-called four times means that in order to detect the same target bonding area at different positions, images located up, down, left, and right (upper left, lower right, lower left, and lower right) based on the visual center are acquired.

In the case of the bonding area 110' located on the visual center line, the deviation value according to the working position is small, so an accurate position value can be obtained, but the bonding area 110' located at each periphery may produce mechanical, optical Therefore, when the same adhesive region 110' is arranged at different positions (upper left, lower right, lower left and lower right), multiple images acquired from each position are overlapped to obtain corresponding Accurate location information of the target bonding area. In addition, a plurality of images detected at different positions in the target bonding area can be acquired. Therefore, by moving and acquiring images at pitch intervals calculated from the matrix information of the bonded areas, the inspection time can be shortened, and the number of units per hour (UPH) can be increased. fine location judgment.

On the other hand, the method of calculating the pitch in the second embodiment of the present invention is as follows. Usually, fov detects an image in the form of a square circle, and the material detected by fov can obtain an image of M rows×N columns according to the shape of the material.

In this case, M and N may be integers, even numbers, or odd numbers.

When M is an even number, the image of the target bonding area can be obtained by moving M/2 column intervals, and when M is an odd number, the image of the target position can be obtained by moving (M+1)/2 column intervals. Similarly, when N is an even number, the image of the target bonding area can be obtained by moving N/2 line intervals, and when N is an odd number, it can be obtained by moving (N+1)/2 line intervals Image of the targeted bonding area.

In the second embodiment of the present invention, in the case of 3 rows×3 columns, each is an odd number of 3, so the image of the target bonding area can be obtained by moving (3+1)/2=2 pitches.

Similarly, in the case of 4 rows×4 columns, images can be obtained every 2 pitches along with the movement, and in the case of 5 rows×5 columns, images can be obtained every 3 pitches along with the movement.

For reference, depending on the shape of the material, the number of M rows and the number of N columns detected by the visual viewing angle may be different from each other. That is, in the case of detecting an image of 4 rows×3 columns, the column interval is detected by moving every 4/2=2 pitches because M is an even number, and the row interval is moved by every (3 pitches) because N is an odd number. +1)/2=2 pitch to obtain overlapping images.

On the other hand, as shown in FIG. 9, in the semiconductor material attaching device according to the second embodiment of the present invention, before the bonding of the visual unit idles, the objects according to the interval of the photographed bonding area 110' are bonded. For the image of overlapping regions, the vision unit 200 ′ moves the pitch in the X-axis direction at a pitch interval of 2 calculated based on the matrix information of the adhesive region 110 ′.

In the second embodiment of the present invention, an image of 3 rows × 3 columns is obtained in one viewing angle fov. Since M is an odd number, the image is obtained by moving with a pitch of (3+1)/2=2 columns . After moving to the X-axis direction (right side) and obtaining all the images of 1~3 rows relative to the wafer, since N is an odd number, the worktable moves to the Y-axis direction ( Bottom) After moving 2 pitches, move the vision camera to the X-axis direction (left side), and move the pitch at 2-column intervals to obtain images of the bonded area of 3~5 rows relative to the wafer. This process is repeated sequentially, and when images of the bonding regions of the wafer are collected, overlapping images can be obtained for each target bonding region.

For reference, in the first and second embodiments of the present invention, for ease of description, the description is based on the target bonding area located in row 3, but in order to check the target bonding area located in row 1 and row 2, it is also Shooting can be started from the surrounding area where there is no adhesive area 110 ′, so that the images of the target adhesive area located in the 1st and 2nd rows can be obtained based on the center of vision at the upper left, lower right, lower left, and lower right.

The attachment method of the semiconductor material attachment device of the present invention utilizes the visual unit 200' to judge the position error of the adhesive region 110' of the miniaturized semiconductor material, and makes the position error be obtained during the attachment process. To modify, so that the subsequent processes of the attaching process, for example, the bonding process of semiconductor materials or the sputtering process, etc., minimize the defect of the product and the like.

That is, in Embodiment 1 and Embodiment 2 of the present invention, the bonding area of each wafer and the through hole of the tape were inspected as an example before the semiconductor material was attached to the wafer or tape. However, the semiconductor wafer or material is attached to each After the bonding area, whether the adhesion is good or not is checked by the same method in the inspection of the automatic post welding inspection (PBI, post bonding inspection), which can improve the reliability of the inspection.

In this description, the preferred embodiment of the present invention has been described with reference to, but those of ordinary skill in the art to which the present invention belongs will not deviate from the idea and area of the present invention described in the scope of protection of the invention described below. Various modifications and changes can be made to the present invention within the scope. Therefore, as long as the modified implementation basically includes the structural elements of the claimed protection scope of the present invention, it shall be deemed to be included in the technical scope of the present invention.

100‧‧‧Attachment object (band)

110‧‧‧Attachment position (through hole)

200‧‧‧visual unit

fov‧‧‧perspective

Claims (12)

A method of attaching a semiconductor material, which is a circuit substrate having a bonding area formed with a plurality of semiconductor materials for bonding, a workbench for placing the circuit substrate, and a vision unit for photographing the bonding area of the circuit substrate The attachment method of the semiconductor material attachment device, comprising: using the above-mentioned vision unit to take a single-lens shot of the target bonding area to be bonded with the above-mentioned semiconductor material and the first shooting step of a plurality of adjacent bonding areas; After the above-mentioned target bonding area, in order to make the target bonding area enter other positions within the viewing angle of the above-mentioned visual unit, the step of moving the visual unit or the workbench at an interval calculated based on the matrix information of the bonding area entering the visual angle of the visual unit ; In the state of transferring the vision unit or the workbench at the above-mentioned interval calculated, use the above-mentioned vision unit to photograph the second shooting step of the above-mentioned target bonding area; by repeating the above-mentioned step of transferring the vision unit or workbench and the second Two photographing steps, a step of acquiring a plurality of images of the above-mentioned target bonding area in different positions within the viewing angle of the above-mentioned visual unit; and judging the above-mentioned target from the acquired images of the plurality of above-mentioned target bonding areas Steps for the location of the bonded area. A method for attaching a semiconductor material, which has a through hole formed with a plurality of bumps for accommodating the semiconductor material and for the above-mentioned semiconductor The sputtering process of the material is attached to the tape of the template, the table for placing the above tape and the attachment method of the semiconductor material attachment device of the vision unit for photographing the through hole of the above tape, comprising: using the above vision unit, The first photographing step of shooting the target through-hole and a plurality of adjacent through-holes to accommodate the bumps of the semiconductor material with a single lens; In other positions, the step of transferring the vision unit or workbench at an interval calculated according to the matrix information of the through hole of the viewing angle entering the vision unit; in the state of moving the vision unit or workbench at the above-mentioned calculated interval, using the above-mentioned vision unit The second photographing step of photographing the above-mentioned target through-hole; by repeating the above-mentioned step of transferring the vision unit or the workbench and the second photographing step several times, the position of the above-mentioned target through-hole in different positions can be acquired within the viewing angle of the above-mentioned vision unit a step of a plurality of images; and a step of judging the position of the target through hole from the acquired images of the plurality of target through holes. The semiconductor material attachment method according to claim 1 or 2, characterized in that the above-mentioned first photographing step and the second photographing step are repeated multiple times at respective positions to photograph, and the above-mentioned The step of position is carried out by repeating multiple times at the respective positions The position is judged by the average value of a plurality of position values obtained in the first photographing step and the second photographing step. According to the semiconductor material attaching method described in claim 1 or 2, it is characterized in that the above-mentioned first photographing step and the second photographing step are repeated multiple times at their respective positions and photographed, and obtained according to the above-mentioned repeated multiple times of photographing The step of judging the position is as follows: when a specific abnormal position value is found among the plurality of position values obtained in the first photographing step and the second photographing step by repeating the respective positions multiple times, filter Corresponding data, when different deviations occur in multiple data among the multiple position values obtained in the above-mentioned first shooting step and second shooting step that are repeated many times, recalibrate or change the corresponding position value is considered bad. According to the semiconductor material attachment method described in claim 1, it is characterized in that the bonding area entering the viewing angle of the above-mentioned visual unit is formed as M rows×N columns, and the above-mentioned M and N are integers, and in the above-mentioned second photographing step, When the above-mentioned M is an even number, the above-mentioned target bonding area is photographed while moving the M/2 column interval, and when the above-mentioned M is an odd number, the above-mentioned target bonding area is photographed while moving (M+1)/2 column intervals, When the above-mentioned N is an even number, the above-mentioned target bonding area is photographed while moving N/2 line intervals, and when the above-mentioned N is an odd number, the above-mentioned target bonding area is photographed while moving (N+1)/2 line intervals. According to the semiconductor material attachment method described in Claim 2, it is characterized in that the through hole entering the viewing angle of the above-mentioned visual unit is formed as M rows×N columns, and the above-mentioned M and N are integers. In the above-mentioned second shooting step, when When the above-mentioned M is an even number, the above-mentioned target through-hole is photographed while moving the interval of M/2 columns; When it is an even number, the above-mentioned target through hole is photographed while moving N/2 line intervals, and when the above-mentioned N is an odd number, the above-mentioned target through hole is photographed while moving (N+1)/2 lines. The method for attaching a semiconductor material according to claim 1, wherein the second photographing step is: while moving a pitch interval with the vision unit or the workbench, use the vision unit to photograph the target adhesion area. The method for attaching semiconductor materials according to Claim 2, wherein the second photographing step is: photographing the target through-hole with the vision unit while moving a pitch interval with the vision unit or the workbench . The method for attaching a semiconductor material according to claim 1, wherein in the step of acquiring a plurality of images of the target bonding area at different positions, in order to obtain the target bonding area with the above-mentioned visual The center of the unit is the interval calculated based on the images located at the upper left, upper right, lower left and lower right, and the vision unit is used to capture images. The method for attaching a semiconductor material according to claim 2, wherein in the step of acquiring a plurality of images of the target through hole at different positions from each other, in order to obtain the target through hole with the vision unit Using the distance calculated from the images located at the upper left, upper right, lower left, and lower right with the center as the reference, the vision unit or the workbench is moved while the vision unit is used to capture images. The method for attaching a semiconductor material according to claim 2, wherein the template has a There are a plurality of through-holes larger than the through-holes of the above-mentioned tape. In the above-mentioned first photographing step or the above-mentioned second photographing step, the above-mentioned The tolerance between the through hole of the template and the through hole of the above belt, and confirm whether the obtained tolerance is within the initial setting range. The method for attaching a semiconductor material according to claim 2, wherein the template has a plurality of through holes larger than the through holes in the tape in positions corresponding to the through holes formed in the tape, and the judgment target pass After the step of positioning the holes, further comprising: a step of accommodating the bumps of the semiconductor material in the through holes of the tape according to the judged position; and, after accommodating the bumps of the semiconductor material in the through holes of the tape The step of inspecting the adhesion state of the semiconductor material by comparing the tolerance between the through hole of the template and the through hole of the tape with the position of the bump of the semiconductor material.

Images